目的由于夜间图像具有弱曝光、光照条件分布不均以及低对比度等特点,给基于夜间车辆图像的车型识别带来困难。此外,夜间车辆图像上的车型难以肉眼识别,增加了直接基于夜间车辆图像的标定难度。因此,本文从增强夜间车辆图像特征考虑,提...目的由于夜间图像具有弱曝光、光照条件分布不均以及低对比度等特点,给基于夜间车辆图像的车型识别带来困难。此外,夜间车辆图像上的车型难以肉眼识别,增加了直接基于夜间车辆图像的标定难度。因此,本文从增强夜间车辆图像特征考虑,提出一种基于反射和照度分量增强的夜间车辆图像增强网络(night-time vehicle image enhancement network based on reflectance and illumination components,RIC-NVNet),以增强具有区分性的特性,提高车型识别正确率。方法RIC-NVNet网络结构由3个模块组成,分别为信息提取模块、反射增强模块和照度增强模块。在信息提取模块中,提出将原始车辆图像与其灰度处理图相结合作为网络输入,同时改进了照度分量的约束损失,提升了信息提取网络的分量提取效果;在反射分量增强网络中,提出将颜色恢复损失和结构一致性损失相结合,以增强反射增强网络的颜色复原能力和降噪能力,有效提升反射分量的增强效果;在照度分量增强网络中,提出使用自适应性权重系数矩阵,对夜间车辆图像的不同照度区域进行有区别性的增强。结果在模拟夜间车辆图像数据集和真实夜间车辆图像数据集上开展实验,从主观评价来看,该网络能够提升图像整体的对比度,同时完成强曝光区域和弱曝光区域的差异性增强。从客观评价分析,经过本文方法增强后,夜间车型的识别率提升了2%,峰值信噪比(peak signal to noise ratio,PSNR)和结构相似性(structural similarity,SSIM)指标均有相应提升。结论通过主观和客观评价,表明了本文方法在增强夜间车辆图像上的有效性,经过本文方法的增强,能够有效提升夜间车型的识别率,满足智能交通系统的需求。展开更多
Poor illumination greatly affects the quality of obtained images.In this paper,a novel convolutional neural network named DEANet is proposed on the basis of Retinex for low-light image enhancement.DEANet combines the ...Poor illumination greatly affects the quality of obtained images.In this paper,a novel convolutional neural network named DEANet is proposed on the basis of Retinex for low-light image enhancement.DEANet combines the frequency and content information of images and is divided into three subnetworks:decomposition,enhancement,and adjustment networks,which perform image decomposition;denoising,contrast enhancement,and detail preservation;and image adjustment and generation,respectively.The model is trained on the public LOL dataset,and the experimental results show that it outperforms the existing state-of-the-art methods regarding visual effects and image quality.展开更多
文摘目的由于夜间图像具有弱曝光、光照条件分布不均以及低对比度等特点,给基于夜间车辆图像的车型识别带来困难。此外,夜间车辆图像上的车型难以肉眼识别,增加了直接基于夜间车辆图像的标定难度。因此,本文从增强夜间车辆图像特征考虑,提出一种基于反射和照度分量增强的夜间车辆图像增强网络(night-time vehicle image enhancement network based on reflectance and illumination components,RIC-NVNet),以增强具有区分性的特性,提高车型识别正确率。方法RIC-NVNet网络结构由3个模块组成,分别为信息提取模块、反射增强模块和照度增强模块。在信息提取模块中,提出将原始车辆图像与其灰度处理图相结合作为网络输入,同时改进了照度分量的约束损失,提升了信息提取网络的分量提取效果;在反射分量增强网络中,提出将颜色恢复损失和结构一致性损失相结合,以增强反射增强网络的颜色复原能力和降噪能力,有效提升反射分量的增强效果;在照度分量增强网络中,提出使用自适应性权重系数矩阵,对夜间车辆图像的不同照度区域进行有区别性的增强。结果在模拟夜间车辆图像数据集和真实夜间车辆图像数据集上开展实验,从主观评价来看,该网络能够提升图像整体的对比度,同时完成强曝光区域和弱曝光区域的差异性增强。从客观评价分析,经过本文方法增强后,夜间车型的识别率提升了2%,峰值信噪比(peak signal to noise ratio,PSNR)和结构相似性(structural similarity,SSIM)指标均有相应提升。结论通过主观和客观评价,表明了本文方法在增强夜间车辆图像上的有效性,经过本文方法的增强,能够有效提升夜间车型的识别率,满足智能交通系统的需求。
基金This work was supported by the Shanghai Aerospace Science and Technology Innovation Fund(No.SAST2019-048)the Cross-Media Intelligent Technology Project of Beijing National Research Center for Information Science and Technology(BNRist)(No.BNR2019TD01022).
文摘Poor illumination greatly affects the quality of obtained images.In this paper,a novel convolutional neural network named DEANet is proposed on the basis of Retinex for low-light image enhancement.DEANet combines the frequency and content information of images and is divided into three subnetworks:decomposition,enhancement,and adjustment networks,which perform image decomposition;denoising,contrast enhancement,and detail preservation;and image adjustment and generation,respectively.The model is trained on the public LOL dataset,and the experimental results show that it outperforms the existing state-of-the-art methods regarding visual effects and image quality.
